THE CHALLENGE OF DIAGNOSIS

DXA in Adults
In adults, the advent of noninvasive bone densitometry has offered a means to identify
and treat individuals with bone fragility before they fracture and to monitor their response
to therapy with parameters other than fracture. Of the available densitometry techniques,
DXA is currently the preferred method for detecting adults at risk for osteoporosis. Its
widespread use as a clinical tool is in part because of its low radiation exposure, excellent
precision, ease of testing, and affordability.
For postmenopausal women, much work has been done to establish disease severity
thresholds and even fracture risk based on DXA. In this group, fracture risk has been
correlated to low bone density as measured by DXA. The World Health Organization has
developed criteria for the diagnosis of osteoporosis in postmenopausal Caucasian
women based on a BMD that is 2.5 standard deviations or more below the average value
for a young adult (i.e., T-score < 2.5).
A significant limitation of DXA is that while it uses density as a surrogate for strength,
it does not truly measure all parameters of bone that determine fracture risk. For example,
the reduction in fracture rates observed after initiation of bisphosphonate therapy exceeds
that predicted by gains in BMD (7981). This observation illustrates the importance of
factors other than bone mineral that contribute to bone strength. The size, shape, geometry,
microarchitecture of bone and the rates of bone turnover are important modifiers of
bone strength and fracture risk.
In contrast to that of postmenopausal women, the diagnosis of bone fragility in men,
younger women, and especially children is more complex and controversial (82). The
indications for bone DXA in these patients and the clinical implications of their results
are still being debated. Experts in the bone field have proposed guidelines for testing and
interpreting DXA results in men, young women, and children, based on opinion where
data were lacking (83). Another panel of bone experts have criticized these recommendations,
citing the lack of objective data to support the opinions (84). In short, considerable
controversy surrounds the optimal approach to identify risk for bone fragility in men
and younger individuals.

DXA in Children and Adolescents

Children present the most challenging population for assessing skeletal health, primarily
because of the numerous variables of growth. Measurement techniques in the pediatric
setting would ideally be safe, painless, of short duration, and would provide valuable
information. In comparison with other bone measurement systems, DXA best fits these
criteria. However, there are difficulties that are unique to children and adolescents when
using this tool. This discussion serves as a very brief introduction to topics that are the
focus of this text.
Bones change in size, shape, and mass throughout the first two decades of life, and the
tempo of change varies by skeletal site and individual. Measurements of bone mass by
DXA are two-dimensional (i.e., areal) and are strongly influenced by bone size, pubertal
stage, and bone age (85,86). Children with smaller bones may appear to have low BMD,
and, in serial testing, changes resulting from increased bone size can be misconstrued for
increased bone density. For this reason, areal BMD can be a source of confusion in the
pediatric population, and the concept of volumetric density may be more appropriate.
When using DXA with children, it may be that different units of measurement will be
more useful than those for adults. For example, BMC and BMD are often used interchangeably
to denote mass, although they are very different parameters. It appears that
BMC measured by DXA is more sensitive to change in bone acquisition than is areal
BMD, especially in early- and prepubertal children (86).

Another difficulty encountered in the use of DXA is the lack of universal pediatric
reference data for determining normal from abnormal bone mass. Until recently, DXA
software programs automatically generated a T-score, comparing the data of the subject,
regardless of his or her age, with that of healthy young adults. This is an inappropriate
comparison for those under age 20 who have not yet achieved PBM.
Even when comparing children to their age- and gender-matched peers, there is difficulty
because the tempo of growth, sexual maturation, and bone mineral accrual can
vary among individuals and can be altered by chronic illness. These factors must be
considered as well in determining if bone mineral is normal.
The complexity of obtaining and interpreting bone densitometry in children and
adolescents has led to confusion and misdiagnoses in children. In one recent study, more
than half of the subjects referred for a evaluation of pediatric osteoporosis had been
misdiagnosed with low bone mass, with the most frequent error resulting from the use of
a T-score in pediatric patients (87). As this is a developing field, DXAs are frequently
performed and interpreted by specialists with expertise in adult osteoporosis but with
limited experience with pediatric densitometry. Misleading information about bone mass
can result from the use of inappropriate software or improper positioning during acquisition,
as well as from an interpretation of results that does not account for known confounding
variables.
The consequences of these errors can be costly. Pediatric patients may be inappropriately
labeled as osteoporotic, producing anxiety in parents and children. Physicians
may respond to these reports by restricting physical activity or by prescribing drugs for
osteoporosis that are, to date, untested for safety and efficacy in children. In addition, if
the results of these studies are confusing or are thought to be unreliable, the clinician is less likely to
initiate skeletal health assessment in children using DXA, missing an
opportunity to identify and correct deficits in developing bone.

SUMMARY
There is an ever-expanding body of knowledge regarding the positive and negative
influences on developing bone. Despite this, we are observing worrisome trends in childhood
such as poor nutrition, sedentary lifestyle, and obesity, all of which are associated
with low bone mass. Beyond that, increasing rates of childhood fractures have been
reported on several continents. In addition, more children are surviving significant illnesses
and treatment regiments that can have profound deleterious effects on bone.
Aside from immediate concerns in children, the model of PBM as a determinant of
adult osteoporosis and fragility fracture implies that the first two decades of life represent
a window of opportunity in which to implement upstream prevention and intervention
strategies that may impart enduring effects on the bone health of an individual.
It follows that this same period represents a window of vulnerability and a time
during which increased scrutiny of bone development is essential.
It is therefore critical that we expand our ability to measure bone health parameters in
the growing patient, to identify markers of inadequate gain, and to monitor effectiveness
of interventions. This requires a noninvasive, safe, and available instrument with good
precision, short test time, and useful output. As DXA, even with its limitations, is currently
the best fit for bone health assessment in children, it is imperative that the clinical
utility of DXA be maximized so that we can recognize indications of bone fragility and
identify trajectories of bone acquisition that may predispose a child to a lifetime of poor
bone health.
Development of these guidelines for the clinical use of DXA in pediatric patients will
hopefully improve the quality of densitometry data and reduce the frequency of misdiagnosis
in the clinical setting while research continues to advance the usefulness of this
and other tools.

WHAT ARE WE MEASURING WITH BONE DENSITOMETRY?

Bone densitometry offers a tool with which pediatric bone status can be assessed. As
the child grows, the skeleton will increase in size and mineral content and will change in
shape. When interpreting measurements from bone densitometry scanners, it is imperative
that these changes in bone size, shape, and mass are taken into account (20,21). For example, changes in
bone density over time could reflect changes in bone size, mineral
content, or a combination of these. The quantitative measures that can be obtained from
most densitometry techniques include bone area (BA; cm2), bone mineral content (BMC; g),
and bone mineral density (BMD; g/cm2).
A model based on the biological organization of bone was proposed by Rauch and
Schonau (22) to help in understanding and interpreting the measurements obtained from
bone densitometry and to relate these changes to the physiological changes that occur
during bone development (see Fig. 1). The model describes separate definitions for the
material, compartment, and total densities of bone, and each of these will be discussed
briefly:
1. Material mineral density. This reflects the degree of mineralization of the organic bone
matrix. Material density can be determined only within a very small volume occupied
only by bone matrix, exclusive of marrow spaces, osteonal canals, lacunae, and canaliculi.
The resolution required to measure BMDmaterial is not possible with current
noninvasive densitometric techniques; BMDmaterial can be determined from specimens
taken at bone biopsy, an invasive procedure. These specimens can be analyzed by mineral/
ash weight, contact radiography, backscatter electron microscopy, or laser-ablated
mass spectrometry. Measurement of BMDmaterial is not routinely assessed in clinical
practice.
2. Compartment mineral density. The BMDcompartment is the amount of mineral contained
within the trabecular or cortical compartments (i.e., the mass of mineral per unit volume
of trabecular or cortical bone). Quantitative computed tomography (QCT) measures
cortical and trabecular bone separately, and can therefore measure BMD compartment in both
types of bone. DXA measurements are a composite of trabecular and cortical bone, and
so the technique is not able to separate the two components at most sites. BMD compartment
can be determined by DXA in skeletal sites such as the diaphyses of the femur and the
radius, both of which are comprised of cortical bone. Radiogrammetry measures the
cortical BMDcompartment of the metacarpals.
3. Total mineral density. BMDtotal is the mineral density of all of the material contained
within the periosteal envelope and articular surfaces. QCT and DXA measure BMD total.
Calculations are required to estimate bone volume from DXA scans because this technique
measures areal density only. Bone mineral apparent density (BMAD) is an example of a volumetric density
calculated using BMDtotal. This density is sometimes inappropriately
referred to in the literature as true bone density.
Table 3 summarizes which of the aforementioned BMD measurements can be determined
using the densitometry techniques discussed in this chapter; quantitative ultrasound
(QUS) and magnetic resonance imaging (MRI) do not measure BMD

COMPARISON BETWEEN CENTRAL AND PERIPHERAL

TECHNIQUES
In adults, measurements of bone mass at both axial and peripheral sites have proven
to predict future osteoporotic fracture (128). In older adults, osteoporosis is defined in
terms of bone densitometry as a T-score (i.e., the SD from the mean of ethnic- and sexmatched
peak BMD) of 2.5 or below using axial DXA in the lumbar spine and proximal
femur. The agreement in classification by the various densitometric techniques has been
studied in adults (129131), and each performs well in differentiating osteoporosis or
osteopenia from normal bone status. However, each technique identifies different people

as osteoporotic or osteopenic; hence, the diagnostic agreement among the methods is poor (i.e., a * score
of 0.4). As an exception to the rule, several studies have shown that
the agreement between trabecular vBMD (measured by QCT) and lateral DXA BMD has
a score of 0.75 (131). The reasons for poor agreement among different bone density
methods and at different sites are likely to include differences in the ability of the technique
to measure integral or separate cortical and trabecular bone (132,133), differing
patterns of regional bone loss (e.g., in the spine versus the radius), and differential diseasespecific effects on bone. Differences in scanner technology will also be relevant in
contributing to the poor agreement among methods. Whether BMD is measured in adults
or in children, the agreement among different techniques is likely to be of similar magnitude
(r between 0.4 and 0.9).
Because there may be regional differences in bone mass and strength, selection of
skeletal site to scan is important. For example, in children with juvenile idiopathic arthritis,
who are most likely to suffer a vertebral crush fracture (134,135), measurement of
spinal trabecular bone should be a priority. Any measurement that does not include the
spine is less likely to be sensitive to the bone changes that occur. Diagnostic agreement
between axial and peripheral skeletal sites may also differ depending on the childs phase
of skeletal development. A large change in DXA spinal BMD with no change in radius
trabecular BMD may be caused by the increase in bone size due to the pubertal growth
spurt rather than being due to the change in volumetric bone mineral density. The relationship
between the peripheral and axial bone densitometry techniques and fractures has
not been studied in children.
Several studies have been performed that investigate the ability of peripheral measurement
to predict osteoporotic fracture in adults (128,136,137). Site-specific measurements
have proven to be the best predictors of fractures at that site; for example, hip BMD
will predict hip fracture better than radial or spinal BMD measurements. However, BMD
measurements by peripheral techniques do predict spine and hip fracture in adults, thus
providing useful information if an axial BMD measurement is not available.
The forearm is the most common site of fracture in children. Goulding et al. (138) have
shown that children who have had fractures generally have lower BMD in the whole
skeleton. Some studies have confirmed an association between low BMD and all upper
limb fractures (139), whereas others have observed reductions in hip and spine but not
whole-body bone measurements in children who have fractured (140,141). In the only
prospective study of childhood fracture to date, low BMD, as measured by axial DXA,
was predictive of the likelihood of a child to refracture within 4 yr of the initial fracture
date (142). The correlation between BMD and childhood fractures has been reviewed
(143). In young people, lower bone density at the spine or whole body has been linked
to fractures only at the forearm but not at other skeletal sites. These findings suggest that
low BMD may be a contributing factor to childhood fracture, just as it is in adults.
However, there are insufficient data to establish a fracture threshold in children and
young adults. Furthermore, comparisons among different scanning techniques for childhood
fractures have not yet been made.